List your company

Reinforcement Learning on Minepump 1.0 (evaluation)

-669.16RewardDQN + Static 1
-1,032.2552-937.9901-843.725-749.4599Nov 4, 2025

Evaluation results

MethodLinks
DQN + Static 1Shielding=Static 12025.11-669.1600.01
DQNShielding=None2025.11-672.360—
DQN + Static (*)Shielding=Static (*)2025.11-74700.4
DQN + AdaptiveShielding=Adaptive2025.11-764.9810.19
DQN + Static 2Shielding=Static 22025.11-813.900.4
Static Ctrl. 1Type=Symbolic Controller2025.11-821.10—
Static Ctrl. 2Type=Symbolic Controller2025.11-1,018.291—